Unexpected GEOFAR tracking errors

Dear community

My name is David and I try to simulate the spectral gamma-ray detector response of a scintillation detector. For that puprose, I use three custom user-routines as well as a quite complex mass model/geometry. The entire framework was successfully tested and validated with experimental data for simplified mass models.

With the more complex mass model, I’ve encountered now some fatal GEOFAR errors for some runs (I run my simulations in parallel runs on a cluster). Please refer to the attached error and log files. I’m aware that these GEOFAR errors might indicate geometrical or numerical precision errors in the input file as shown in previous posts (Geofar error when the beam energy is increased, Gamma Reactions

Problem Solving so far
To check if there are any geometrical errors in my input script, I’ve checked the end points indicated in the error files manually in FLAIR. In addition, I’ve run the debugger in flair for the entire geometry as well as for a smaller region around the problematic regions. To my surprise, no geometrical error was found for both cases…

In my opinion, these errors might therefore indicate numerical precision errors like reported in Gamma Reactions. This is supported by the fact that the errors occured in multpiple regions across my geometry often very close to the boundaries between regions with small dimensions. However, the geometry isn’t that large in my case (<5E3cm) with the smallest geometries having a dimension >0.1mm. In contrast to Gamma Reactions, I got also fatal errors.


  1. If not geometrical errors, might these errors indicate numerical precision errors?
  2. If yes, how can I prevent these errors?
  3. Might it be beneficial to change the accuracy or the parenthesis options in the geobegin card?

Please note that only the first couple of errors are contained in the error file to ensure the file size is small enough for sharing here. In addition, these attached files are only for one run. I’ve other run files with similar errors but in different regions.

Please further note that, due to publication restrictions, I can’t share the input-file and the user-routines publically. However, if you want to have a look at these files, I’m happy to share them with you by e-mail or by the private section in this forum. Please contact me in the personal fluka section or by e-mail (david.breitenmoser@psi.ch)

design00001_run077_001.log (162 Bytes)
randesign00001_run077_002.txt (1.6 KB)
randesign00001_run077_001.txt (1.6 KB)
design00001_run077_001.err (411.8 KB)

Additional Notes
I’m using FLUKA 4-2.2 and FLAIR 3.1-15.1.

Thank you very much for your support
Kind regards,

Dear @BV96

many thanks for reporting the issue. With the help of your input files we found a legacy bug in the handling of the QUA, as well a couple of small bugs in flair. The fix(s) will appear with the next fluka minor release

best regards

1 Like